Have you noticed that we don’t see so many articles these days about how autonomous vehicles will be taking over the automotive industry and that there will be no need for parking in the future? What has happened to the AV industry?
There is an article Astrid put up last week on Parknews.biz titled “Driverless Cars Show the Limits of Today’s AI” that may explain it. The article, from the Economist, posits that artificial intelligence, particularly in dealing with complex issues like driving, has a long way to go. Problems like dealing with an airplane landing on a highway or someone jumping out in a chicken suit or a stop sign covered with stickers, often flummox the computer, but is something a human can deal with without thinking.
Computers learn through the invocation of Moore’s Law. Computing power gets cheaper and faster on a exponentially. In other words, it processes terabytes of data (which it must do to drive a car) using brute force. Its not smart, just fast.
IN MARCH Starsky Robotics, a self-driving lorry firm based in San Francisco, closed down. Stefan Seltz-Axmacher, its founder, gave several reasons for its failure. Investors’ interest was already cooling, owing to a run of poorly performing tech-sector IPOs and a recession in the trucking business. His firm’s focus on safety, he wrote, did not go down well with impatient funders, who preferred to see a steady stream of whizzy new features. But the biggest problem was that the technology was simply not up to the job. “Supervised machine learning doesn’t live up to the hype. It isn’t actual artificial intelligence akin to c-3PO [a humanoid robot from the “Star Wars” films]. It’s a sophisticated pattern-matching tool.”
Wow! VC is impatient, tech simply is not up to the job. Who knew?
One study, for instance, found that computer-vision systems were thrown when snow partly obscured lane markings. Another found that a handful of stickers could cause a car to misidentify a “stop” sign as one showing a speed limit of 45mph. Even unobscured objects can baffle computers when seen in unusual orientations: in one paper a motorbike was classified as a parachute or a bobsled. Fixing such issues has proved extremely difficult, says Mr Seltz-Axmacher. “A lot of people thought that filling in the last 10% would be harder than the first 90%”, he says. “But not that it would be ten thousand times harder.”
This is what happens when people begin to believe their own press. So called “journalists,” or tech writers, simply want the story to be true. They don’t ask the hard questions, they don’t press for answers. After all, if they can do it in Star Wars, why not in 21st century America.
Elon Musk has discovered that its easier to fire a Tesla into orbit, or shuttle humans to the international space station than to have a Tesla drive itself in a snow storm. The technocrats are finding that wishing it so is easier than making it so.
As I have said before, but it bears repeating. Self Driving vehicles will first be shuttles in a confined environment, then perhaps long haul trucking and Grubhub type deliveries, and then taxis in a limited geographic area. Level 5 AV, the ‘Jetson’ solution, is a long time off, if ever, with current technology.